Patents Assigned to Microchip Technology, Inc.
-
Publication number: 20220383970Abstract: A method for outlier management at a flash controller includes testing a flash memory device to identify one or more outlier blocks of the flash memory device. Hyperparameters for a DNN are loaded into a training circuit of the flash controller. Test reads of the one or more outlier blocks are performed and a number of errors in the test reads is identified. The DNN is trained using a mini-batch training process and using the identified number of errors in the test reads and is tested to determine whether the trained DNN meets a training error threshold. The performing, the identifying, the training and the testing are repeated until the trained DNN meets the training error threshold to identify parameters of an outlier-block DNN. A neural network operation is performed using the identified parameters to predict a set of TVSO values. A read is performed using the set of predicted TVSO values.Type: ApplicationFiled: October 21, 2021Publication date: December 1, 2022Applicant: Microchip Technology Inc.Inventors: Lorenzo Zuolo, Rino Micheloni
-
Publication number: 20220382688Abstract: System and method for analyzing CXL flits at read bypass detection logic to identify bypass memory read requests and transmitting the identified bypass memory read requests over a read request bypass path directly to a transaction/application layer of the CXL memory controller, wherein the read request bypass path does not include an arbitration/multiplexing layer and a link layer of the CXL memory controller, thereby reducing the latency inherent in a CXL memory controller.Type: ApplicationFiled: May 10, 2022Publication date: December 1, 2022Applicant: Microchip Technology Inc.Inventors: Sanjay GOYAL, Larrie Simon Carr, Patrick Bailey
-
Publication number: 20220382945Abstract: A method and apparatus for estimating signal related delays in a PLD design is disclosed. The PLD design is modeled in relation to one or more stages, each of the stages including a driver and one or more receivers coupled to the driver with a wiring tree. The modeling is based on a selected set of parameters that include: slope related delays associated with the driver; a delay related to a layout of the wiring tree; and a parameter related to a slope transfer from a previous driver input. A predetermined set of values for each of the selected parameters are accessed; the estimated signal related delays are computed for each of the modeled stages; and are written to a computer-readable storage medium.Type: ApplicationFiled: May 10, 2022Publication date: December 1, 2022Applicant: Microchip Technology Inc.Inventors: Jonathan W. Greene, Gabriel Barajas, Fei Li, Hassan Hassan, James Sumit Tandon
-
Patent number: 11514992Abstract: A method for reading a flash memory device includes storing configuration files of reliability-state Classification Neural Network (CNN) models and Regression Neural Network (RNN) inference models, and storing reliability-state tags corresponding to reliability states. The current number of P/E cycles is identified and a reliability-state CNN model is selected corresponding to the current number of P/E cycles. A neural network operation of the selected reliability-state CNN model is performed to identify a predicted reliability state. Corresponding reliability-state tags are identified and a corresponding RNN inference model is selected. A neural network operation of the selected RNN inference model is performed, using the reliability-state tags as input, to generate output indicating the shape of a threshold-voltage-shift read-error (TVS-RE) curve.Type: GrantFiled: April 20, 2021Date of Patent: November 29, 2022Assignee: Microchip Technology Inc.Inventors: Lorenzo Zuolo, Rino Micheloni
-
Patent number: 11514994Abstract: A method for outlier management at a flash controller includes testing a flash memory device to identify one or more outlier blocks of the flash memory device. Hyperparameters for a DNN are loaded into a training circuit of the flash controller. Test reads of the one or more outlier blocks are performed and a number of errors in the test reads is identified. The DNN is trained using a mini-batch training process and using the identified number of errors in the test reads and is tested to determine whether the trained DNN meets a training error threshold. The performing, the identifying, the training and the testing are repeated until the trained DNN meets the training error threshold to identify parameters of an outlier-block DNN. A neural network operation is performed using the identified parameters to predict a set of TVSO values. A read is performed using the set of predicted TVSO values.Type: GrantFiled: October 21, 2021Date of Patent: November 29, 2022Assignee: Microchip Technology Inc.Inventors: Lorenzo Zuolo, Rino Micheloni
-
Publication number: 20220375532Abstract: A method for performing a read of a flash memory includes storing configuration files for a plurality of RRD-compensating RNNs. A current number of PE cycles for a flash memory are identified and TVSO values are identified corresponding to the current number of PE cycles. A current retention time and a current number of read disturbs for the flash memory are identified. The configuration file of the RRD-compensating RNN corresponding to the current number of PE cycles, the current retention time and current number of read disturbs is selected and is loaded into a neural network engine to form an RNN core in the neural network engine. A neural network operation of the RNN core is performed to predict RRD-compensated TVSO values. The input to the neural network operation includes the identified TVSO values. A read of the flash memory is performed using the predicted RRD-compensated TVSO values.Type: ApplicationFiled: July 26, 2021Publication date: November 24, 2022Applicant: Microchip Technology Inc.Inventors: Lorenzo Zuolo, Rino Micheloni
-
Publication number: 20220376693Abstract: A logic cell for a programmable logic integrated circuit apparatus includes a K-input lookup table (LUT) circuit having a primary output Y, wherein Y is any function of K inputs, and at least one additional output (F). A carry circuit receives the outputs of the LUT and a carry-in input CI. The carry circuit generates a sum output S and a carry-out output CO. The carry circuit can be configured to provide S=CI and select CO from the set {0, 1, F}. The carry circuit can alternatively be configured to provide S=EXOR(Y, CI) and select CO from the set {0, 1, F}. The carry circuit can alternatively be configured to provide S=EXOR(Y, CI) and CO=CI if Y=q or to select CO from the set {0, 1, F} if Y?q, where q is a pre-determined value (e.g., such as 0 or 1).Type: ApplicationFiled: November 18, 2021Publication date: November 24, 2022Applicant: Microchip Technology Inc.Inventors: Jonathan W. Greene, Marcel Derevlean
-
Publication number: 20220374169Abstract: A method for meeting quality of service (QoS) requirements in a flash controller that includes one or more instruction queues and a neural network engine. A configuration file for a QoS neural network is loaded into the neural network engine. A current command is received at the instruction queue(s). Feature values corresponding to commands in the instruction queue(s) are identified and are loaded into the neural network engine. A neural network operation of the QoS neural network is performed using as input the identified feature values to predict latency of the current command. The predicted latency is compared to a first latency threshold. When the predicted latency exceeds the first latency threshold one or more of the commands in the instruction queue(s) are modified. The commands are not modified when the predicted latency does not exceed the latency threshold. A next command in the instruction queue(s) is then performed.Type: ApplicationFiled: August 10, 2021Publication date: November 24, 2022Applicant: Microchip Technology Inc.Inventors: Lorenzo Zuolo, Rino Micheloni
-
Publication number: 20220342590Abstract: In one implementation a vector processor gather/scatter apparatus comprises a plurality of vector ports, and a random access memory, where the plurality of vector ports are in communication with the random access memory, and where one or more of the plurality of vector ports uses one or more of an address register and one or more of a stride register in communication with the random access memory to allow the gather/scatter of random access memory contents.Type: ApplicationFiled: February 11, 2022Publication date: October 27, 2022Applicant: Microchip Technology Inc.Inventor: Christopher I. W. NORRIE
-
Publication number: 20220342844Abstract: In one implementation a vector processor unit having preload registers for at least some of vector length, vector constant, vector address, and vector stride. Each preload register has an input and an output. All the preload register inputs are coupled to receive a new vector parameters. Each of the preload registers' outputs are coupled to a first input of a respective multiplexor, and the second input of all the respective multiplexors are coupled to the new vector parameters.Type: ApplicationFiled: March 22, 2022Publication date: October 27, 2022Applicant: Microchip Technology Inc.Inventor: Christopher I. W. NORRIE
-
Publication number: 20220342668Abstract: In one implementation devoid of an effective address generator a method of call operation comprises pushing one or more parameters onto a first stack, pushing the contents of one or more registers onto a second stack, popping off the first stack one or more of the parameters into one or more of the registers whose contents were pushed onto the second stack, performing register to register operations on the one or more registers whose contents were pushed onto the second stack with a result of the register to register operations being stored in a result register, the result register being one of the one or more registers whose contents were pushed onto the second stack, popping off the second stack the contents of all the one or more registers into their respective registers from which they came, and returning control to an instruction following the call.Type: ApplicationFiled: September 7, 2021Publication date: October 27, 2022Applicant: Microchip Technology Inc.Inventor: Christopher I. W. NORRIE
-
Publication number: 20220342582Abstract: A system and method for double data rate (DDR) chip-kill decoding using an array of ECC erasure decoders operating in parallel to identify a location of a chip-kill event associated with a DDR memory device. The ECC erasure decoder that correctly identifies the chip-kill location is then used to decode subsequent DDR bursts transmitted from the DDR memory device.Type: ApplicationFiled: February 14, 2022Publication date: October 27, 2022Applicant: Microchip Technology Inc.Inventor: Peter John Waldemar Graumann
-
Publication number: 20220300793Abstract: A method and apparatus for performing a convolution of a N×N matrix. A weights matrix for a N×N Convolutional Neural Network (CNN) is received and is divided into 3×3 weights matrixes. Lines of image values are read and are stored in a buffer as sets of image values. A 3×3 convolution is performed to generate a 3×3 convolution value. All 3×3 convolution values associated with a particular N×N convolution and a particular set of image values are summed. The 3×3 convolutions and the summing are repeated until all columns in the set of image values have been processed; and the reading, the storing, the performing 3×3 convolutions, the summing and the repeating performing 3×3 convolutions are repeated until all lines of image values have been processed. The sums associated with a particular N×N convolution are added together to generate an N×N convolution value for each of the N×N convolutions.Type: ApplicationFiled: September 21, 2021Publication date: September 22, 2022Applicant: Microchip Technology Inc.Inventors: Sathishkumar Donthu, Battu Prakash Reddy
-
Patent number: 11449456Abstract: System and method for sharing a PCIe endpoint device with a plurality of host computers, by allocating a quantum of time to a host computer of a plurality of host computers coupled to a PCIe switch, wherein the quantum of time identifies a duration of time during which the host computer has exclusive access to a shareable PCIe endpoint device coupled to the PCIe switch. Requests from the host computer are transmitted to an emulated PCIe endpoint device of the PCIe switch during the quantum of time and the requests are then redirected from the emulated PCIe endpoint device to the shareable PCIe endpoint device during the quantum of time allocated to the host computer.Type: GrantFiled: January 18, 2021Date of Patent: September 20, 2022Assignee: Microchip Technology Inc.Inventors: Derin Jose, Sachindranath Pv, Viswas G
-
Publication number: 20220270698Abstract: A method for reading a flash memory device includes storing configuration files of reliability-state Classification Neural Network (CNN) models and Regression Neural Network (RNN) inference models, and storing reliability-state tags corresponding to reliability states. The current number of P/E cycles is identified and a reliability-state CNN model is selected corresponding to the current number of P/E cycles. A neural network operation of the selected reliability-state CNN model is performed to identify a predicted reliability state. Corresponding reliability-state tags are identified and a corresponding RNN inference model is selected. A neural network operation of the selected RNN inference model is performed, using the reliability-state tags as input, to generate output indicating the shape of a threshold-voltage-shift read-error (TVS-RE) curve.Type: ApplicationFiled: April 20, 2021Publication date: August 25, 2022Applicant: Microchip Technology Inc.Inventors: Lorenzo Zuolo, Rino Micheloni
-
Patent number: 11424902Abstract: System and method for synchronizing a plurality of nodes to a timing signal using a daisy-chain network having a forward transmission path and a reverse transmission path connected at a midpoint. Latency of the timing signal to the midpoint of the daisy-chain network is determined, a respective latency of the timing signal from the node to the midpoint of the daisy-chain network is determined, and a respective timing offset for each of the plurality of nodes is calculated. A local time-of-day counter at each of the plurality of nodes is adjusted based upon the respective timing offset of the node to synchronize the plurality of nodes to the timing signal.Type: GrantFiled: April 28, 2021Date of Patent: August 23, 2022Assignee: Microchip Technology Inc.Inventors: Atanu Chattopadhyay, Andras de Koos
-
Publication number: 20220262434Abstract: A ReRAM memory array includes ReRAM memory cells and a select circuit having two series-connected select transistors connected in series with a ReRAM device. When ReRAM memory cell(s) are selected for erasing, the bit line coupled to the ReRAM memory cell(s) to be erased is biased at a first voltage potential. The source line coupled to the ReRAM memory cell(s) to be erased is biased at a second voltage potential greater than the first voltage potential, the difference between the first voltage potential and the second voltage potential being sufficient to erase the ReRAM device. The gates of the series-connected select transistors coupled to the ReRAM memory cell(s) to be erased are supplied with positive voltage pulses. The gates of the series-connected select transistors coupled to the ReRAM memory cell(s) unselected for erasing are supplied with a voltage potential insufficient to turn them on.Type: ApplicationFiled: May 4, 2022Publication date: August 18, 2022Applicant: Microchip Technology Inc.Inventors: Victor Nguyen, Fethi Dhaoui, John L. McCollum, Fengliang Xue
-
Patent number: 11398291Abstract: A method and apparatus for determining when actual wear of a flash memory device differs from a reliability state. Configuration files of a reliability-state classification neural network model are stored. The operation of a flash memory device is monitored to identify current physical characteristic values. A read of the flash memory device is performed to determine a number of errors. A neural network operation is performed using as input a set of threshold voltage shift offset values currently being used to perform reads of the flash memory device and the calculated number of errors, to identify a predicted reliability state. The identified current physical characteristic values are compared to corresponding tags associated with the predicted reliability state and a flag or other indication is stored when the comparison indicates that the identified current physical characteristic values do not correspond to the respective tags associated with the predicted reliability state.Type: GrantFiled: March 26, 2021Date of Patent: July 26, 2022Assignee: Microchip Technology Inc.Inventors: Lorenzo Zuolo, Rino Micheloni
-
Patent number: 11392385Abstract: A system and method for monitoring processors operating in lockstep to detect mismatches in pending pipelined instructions being executed by the processors. A lockstep monitor implemented in hardware is provided to detect the mismatches in the pending pipelined instructions executing on the lockstep processors and to initiate an auto-recovery operation at the processors if a mismatch is detected.Type: GrantFiled: October 20, 2020Date of Patent: July 19, 2022Assignee: Microchip Technology Inc.Inventor: Pierre Selwan
-
Publication number: 20220191144Abstract: A network switch and associated method of operation for establishing a low latency transmission path through the network which bypasses the packet queue and scheduler of the switch fabric. The network switch transmits each of a plurality of data packets to the identified destination egress port over the low latency transmission if the data packet is identified to be transmitted over the low latency transmission path from the ingress port to the destination egress port, and transmits the data packet to the destination egress port through the packet queue and scheduler if the data packet is not identified to be transmitted over the low latency transmission path from the ingress port to the destination egress ports.Type: ApplicationFiled: July 23, 2021Publication date: June 16, 2022Applicant: Microchip Technology Inc.Inventor: Morten Terstrup